Perceived risks of COVID-19 in early phase of the pandemic

Fear of COVID-19 was high in Germany during the first phase of the SARS-CoV‑2 pandemic. In a study conducted by the German Socio-Economic Panel (SOEP) between 1 April and 5 July 2020, 5783 participants stated how likely they thought it was that SARS-CoV‑2 would cause them to have a life-threatening illness in the next 12 months (2.3% of all respondents did not give an answer; [1]). In other words, respondents were asked to rate their subjective probability of becoming infected and additionally developing a life-threatening disease on a scale of 0 to 100%. The average subjective probability of a life-threatening infection was 26%. This figure was 25% for men, and 27% for women, and was higher in middle and old age than in younger age (18–29 years: 20%; 31–40 years: 22%; 41–50 years: 27%; 51–60 and 61–70 years: 28%; > 70 years: 26%). Participants with a preexisting disease in 2019 rated their subjective risk slightly higher than participants without such a preexisting disease (27% vs. 23%). Overall, 28% of the respondents estimated their risk to be ≥ 50%.

This fear contrasts strongly with data on COVID-19 during the survey and with figures known 3 years after the onset of the pandemic. This can be seen from the infection fatality rate (IFR), which is the proportion of death cases among the infected:

  • The German Heinsberg Study was conducted after an outbreak of SARS-CoV‑2 in the town of Heinsberg in February 2020. In this study, the COVID-19 IFR was estimated at 0.36% (95% confidence interval: 0.29–0.45; [2]).

  • In 29 mainly high-income countries, recently published age-specific IFRs were as low as 0.002% at 20–29 years, 0.011% at 30–39 years, 0.035% at 40–49 years, 0.123% at 50–59 years, and 0.506% at 60–69 years [3].

On 15 March 2023, the official number of COVID-19 deaths in Germany was 169,097 [4]. If one assumes that for every COVID-19 death there are four life-threatening illnesses, then in a rough calculation there are 280,000 life-threatening illnesses in 1 year. This corresponds to a probability of 0.4% for persons aged 18 years or older. The authors of the SOEP study assumed that 0.6% was the upper limit for the probability of life-threatening SARS-CoV‑2 infections [1].

Although one can argue about the methodology used to estimate IFR in the individual studies and the definition of COVID-19 deaths (deaths due to COVID-19 or deaths with COVID-19), there is no doubt that the risk of contracting life-threatening COVID-19 was massively overestimated in the German population.

This viewpoint offers reasons for the discrepancy between data on COVID-19 and risk perception in Germany. In the following, we show that qualitative attributes of the pandemic, the reporting of the media, and psychological features may have contributed to the overestimation of COVID-19 risks.

Qualitative attributes of the pandemic

The overestimation of the perceived risks of COVID-19 can partly be explained by the qualitative characteristics of the disease [5]. At the beginning of the pandemic, COVID-19 was unknown and perceived as frightening, and people thought that the new pandemic would be difficult to control. Such attributes make a risk seem higher than, for example, the risks of smoking, which may also be high, with thousands of deaths and incident chronic diseases every year. However, risks of smoking are well known, they are taken voluntarily, and they are seen as more controllable (in principle, smokers can quit at any time).

Confusion of basic terms and numbers

German media hardly distinguished between SARS-CoV‑2 test-positive and SARS-CoV-2-infected people. The number of newly infected people was hardly referred to the number of tests. Deaths due to COVID-19 and deaths with COVID-19 (as a comorbidity) were mixed up. For a long time, the Robert Koch Institute reported a case fatality of 4.2% among people with a positive test ignoring the many asymptomatic people who were infected but were never tested. Such inaccuracies made it difficult to have a realistic perception of risk.

Focus on individual cases: denominator neglect

The media tend to favor single cases that stir up emotions, like the story of a young man who had difficulties in walking after infection with COVID-19, and who recovered only slowly. In view of the extremely low probability of young individuals to suffer from such serious consequences—the IFR for the 20–29-year-olds was 0.002% (see above)—such reports strongly mislead the public view on the virus.

The focus on individual cases is related to the neglect of the denominator, which means that much attention is paid to the numerator (the cases that can be counted), while the number of individuals in the population to which the infected cases belong is severely disregarded [6]. For instance, 100 newly infected people may seem impressively large if the population size is not mentioned, because you can imagine each individual—but 0.0001 (100 in 1 million) seems small.

The role of heuristics and confirmation bias

In the current COVID-19 pandemic, the impact of images proved to be extremely powerful. In general, images have a stronger influence on risk perception than numbers do. This was especially true for the images from northern Italy showing military trucks transporting bodies to the crematorium in spring 2020. This is where the availability heuristic comes into play [7].

The consequence of this heuristic is that, for example, the risk of a plane crash is overestimated because of the cognitive availability of television images, whereas the risk of death is underestimated for higher risks, such as dying from a stroke or heart attack, which do not produce dramatic images in the media.

Analogously, the images of patients treated in hospital corridors in northern Italy and New York in 2020 led to an overestimation of COVID-19 risks. Information on the fact that there were only one fifth as many intensive care beds in Italy than in Germany were less cognitively available, and, thus, could not lead to a reduction of subjective risks.

Another heuristic at work is the anchor heuristic. According to Kahneman and Tversky, the first numbers one has in mind—however they came about, however arbitrary they may be—have a major impact on future perceptions [7]. At the beginning of the pandemic, the World Health Organization announced a case-fatality from COVID-19 of 3.4% [8]. From the present perspective, such initial estimates of deaths from COVID-19 were far too high, but may have contributed to greatly exaggerated subjective risks. Once a risk perception—whether high or low—has been made, so-called confirmation bias may become effective. Confirmation bias means that persons prefer to search for information that confirms their hypotheses instead of searching for information that puts their hypotheses into doubt [9]. COVID-19 was far from being a killer virus as initially feared, but confirmation bias made it difficult to correct one’s initial view even if it was strongly exaggerated.

The role of frames

In communication of risk from COVID-19, three types of framing play a role. First, during the pandemic, the view on mortality was restricted to deaths of individuals who died from or with COVID-19. This is framing that excludes collateral effects such as deaths that have to be attributed to measures against COVID-19 (e.g., postponed operations of cancer patients, untreated myocardial infarctions). A second type of framing is the tendency to focus on health alone during the pandemic, and to exclude important aims such as economic growth, employment, the healthy development of children, and social togetherness from the frame. There is a third type of framing: If there is a focus on a specific risk, people have a strong tendency to bring this risk down to zero. Bringing a risk down from 0.5% to 0% may then be seen as a greater success than reducing a risk from 4% to 2%.

Summary

Strong overestimation of the risks of COVID-19 in the early phase of the pandemic was due to characteristics of the disease such as frightfulness and unfamiliarity, to general psychological phenomena such as heuristics and biases, but also to features of media coverage such as framing and stirring up emotions.

Exaggerated fear is not a good advisor. A strong perceived threat of infection shapes behavior, and may reduce gregariousness as well as encourage conformist attitudes [10]. Moreover, a population in fear may be a danger for democracy, and restrictions on fundamental rights may become more likely. We are also aware of the perception that in a pandemic with millions of people infected and many thousands admitted to hospital, thousands of patients also died. Their individual case fatality is tragic.

Conclusion

In a potential future pandemic, people need to be vigilant but not in a panic. To achieve this, some of the factors mentioned in this viewpoint could be taken into account, for example, to contrast the predominance of images with well-prepared figures, and to avoid the denominator neglect instead of using it on purpose. Percentages, which many people find difficult to understand, should be presented graphically or explained by suitable examples. The types of framing presented here should be avoided. Better risk communication can attribute to a more realistic risk perception in future pandemics.